updated collapse feature

woltka/biom.py

@@ -129,6 +129,25 @@ def round_biom(table: biom.Table):
     table.remove_empty(axis='observation')
 
 
+def biom_add_metacol(table: biom.Table, dic, name, missing=''):
+    """Add a metadata column to a table in place based on a dictionary.
+
+    Parameters
+    ----------
+    table : biom.Table
+        Table to add metadata column.
+    dict : dict
+        Metadata column (feature-to-value mapping).
+    name : str
+        Metadata column name.
+    missing : any type, optional
+        Default value if not found in dictionary.
+    """
+    metadata = {x: {name: dic.get(x, missing)} for x in table.ids(
+        'observation')}
+    table.add_metadata(metadata, axis='observation')
+
+
 def collapse_biom(table: biom.Table, mapping: dict, normalize=False):
     """Collapse a BIOM table in many-to-many mode.
 
@@ -158,6 +177,10 @@ def collapse_biom(table: biom.Table, mapping: dict, normalize=False):
     table = table.filter(lambda data, id_, md: id_ in mapping,
                          axis='observation', inplace=False)
 
+    # stop if no feature left
+    if table.is_empty():
+        return table
+
     # add mapping to table metadata
     table.add_metadata({k: dict(part=v) for k, v in mapping.items()},
                        axis='observation')

woltka/cli.py

@@ -241,18 +241,22 @@ def merge_cmd(ctx, **kwargs):
     type=click.Path(exists=True, dir_okay=False),
     help='Path to input profile.')
 @click.option(
-    '--map', '-m', 'map_fp', type=click.Path(exists=True),
+    '--map', '-m', 'map_fp', required=True,
+    type=click.Path(exists=True, dir_okay=False),
     help=('Mapping of lower classification units to higher ones (supports '
           'many-to-many relationships).'))
 @click.option(
     '--output', '-o', 'output_fp', required=True,
     type=click.Path(writable=True, dir_okay=False),
     help='Path to output profile.')
 @click.option(
-    '--normalize', '-n', is_flag=True,
+    '--normalize', '-z', is_flag=True,
     help=('Count each higher classification unit as 1/k (k is the number of '
           'higher classification units mapped to a lower one). Otherwise, '
           'count as one.'))
+@click.option(
+    '--names', '-n', 'names_fp', type=click.Path(exists=True),
+    help='Names of higher classification units to append to output profile.')
 @click.pass_context
 def collapse_cmd(ctx, **kwargs):
     """Collapse a profile based on feature mapping.

woltka/table.py

@@ -21,7 +21,7 @@
 from .file import openzip
 from .biom import (
     table_to_biom, biom_to_table, write_biom, filter_biom, round_biom,
-    collapse_biom)
+    biom_add_metacol, collapse_biom)
 
 
 def prep_table(profile, samples=None, tree=None, rankdic=None, namedic=None,
@@ -440,6 +440,30 @@ def round_table(table):
         del(table[3][i])
 
 
+def table_add_metacol(table, dic, name, missing=''):
+    """Add a metadata column to a table in place based on a dictionary.
+
+    Parameters
+    ----------
+    table : biom.Table, or tuple of (list, list, list, list)
+        Table to add metadata column (data, features, samples, metadata).
+    dict : dict
+        Metadata column (feature-to-value mapping).
+    name : str
+        Metadata column name.
+    missing : any type, optional
+        Default value if not found in dictionary.
+    """
+    # redirect to BIOM module
+    if isinstance(table, Table):
+        biom_add_metacol(table, dic, name, missing='')
+        return
+
+    # add metadata column
+    for feature, metadatum in zip(*(table[1], table[3])):
+        metadatum[name] = dic.get(feature, missing)
+
+
 def collapse_table(table, mapping, normalize=False):
     """Collapse a table by many-to-many mapping.
 
@@ -470,18 +494,17 @@ def collapse_table(table, mapping, normalize=False):
     width = len(samples)
     res = defaultdict(lambda: [0] * width)
     for datum, feature in zip(*table[:2]):
-        try:
-            targets = mapping[feature]
-        except KeyError:
+        if feature not in mapping:
             continue
+        targets = mapping[feature]
         if normalize:
             k = 1 / len(targets)
             datum = [x * k for x in datum]
         for target in targets:
             res[target] = list(map(add, res[target], datum))
 
     # reformat table
-    res = list(res.values()), list(res.keys()), samples, [{}] * len(res)
+    res = list(res.values()), list(res.keys()), samples, [dict() for _ in res]
 
     # round table
     if normalize:

woltka/tests/data/function/go/README.md

@@ -1,3 +1,3 @@
-# UniRef-toGO maps
+# UniRef-to-GO maps
 
 [Gene Ontology](http://geneontology.org/) (**GO**) is a classification system for gene functions. All GO terms can be traced to three top categories: **biological process**, **cellular component**, and **molecular function**. **GO slim** is a set of higher-level categories of GO terms.